Article type
Abstract
Background: We were invited to join a WHO scoping review project involving 4 teams across 6 countries, reviewing the literature on follow-up care of sick and small newborns, covering 1) preterm, SGA and LBW 2) growth faltering in the neonatal period, 3) meningitis and septicaemia 4) hypoxic ischaemia and other serious encephalopathies. Since the included studies might be relevant to more than one of these reviews and we anticipated having a large number to screen, we planned a single 2-stage search, the first stage shared across all teams and the second stage focussing on selecting studies relevant to a team's own review. Here we describe the search workflow and evaluate the functionalities and limitations of using Covidence to manage multiple teams.
Objective: To describe the challenges and successes of using Covidence to select studies across a set of related reviews.
Method and Results: To build the search strategy we utilised a shared Excel sheet which the 4 teams populated with relevant keywords based on their inclusion criteria. A search across five databases was conducted resulting in 25240 articles. We distributed the search results for title and abstract screening equally among the four teams. Teams included any article relevant to any of the four reviews. Since Covidence was not able to allocate the search results equally to four review groups, Endnote Citation Manager was used as an intermediary. This involved importing search results from the five databases into EndNote, removing duplicates, and assigning one-quarter of the total results to each team for abstract screening. Abstracts selected by the four teams were exported from Covidence to Endnote and were combined into one database. This was uploaded four times into Covidence, once for each review team for full-text screening. Full-text articles were bulk uploaded 4 times into each review. Teams reviewed full texts, selected articles for inclusion in their specific review and proceeded to data extraction, (see Figure).
Conclusion: Using Covidence along with a second reference manager is a feasible way of applying a single search to more than one review. Covidence developers could consider adapting their software to allow for similar situations.
Objective: To describe the challenges and successes of using Covidence to select studies across a set of related reviews.
Method and Results: To build the search strategy we utilised a shared Excel sheet which the 4 teams populated with relevant keywords based on their inclusion criteria. A search across five databases was conducted resulting in 25240 articles. We distributed the search results for title and abstract screening equally among the four teams. Teams included any article relevant to any of the four reviews. Since Covidence was not able to allocate the search results equally to four review groups, Endnote Citation Manager was used as an intermediary. This involved importing search results from the five databases into EndNote, removing duplicates, and assigning one-quarter of the total results to each team for abstract screening. Abstracts selected by the four teams were exported from Covidence to Endnote and were combined into one database. This was uploaded four times into Covidence, once for each review team for full-text screening. Full-text articles were bulk uploaded 4 times into each review. Teams reviewed full texts, selected articles for inclusion in their specific review and proceeded to data extraction, (see Figure).
Conclusion: Using Covidence along with a second reference manager is a feasible way of applying a single search to more than one review. Covidence developers could consider adapting their software to allow for similar situations.